Certiorari Petitions and Rehearing En Banc: The Last Two Doors in Federal Appellate Practice

Document type: Article Practice area: Litigation — Appellate Practice Jurisdiction: United States Last reviewed: 5 September 2026

The uncomfortable arithmetic

The Supreme Court receives several thousand petitions for certiorari each Term and grants somewhere in the neighbourhood of sixty. Excluding the in forma pauperis docket, the grant rate among paid petitions is in the low single digits. Courts of appeals grant rehearing en banc in a similarly small fraction of the cases in which it is sought — in most circuits, a handful of cases a year against hundreds of petitions.

The first duty of appellate counsel is therefore to tell the client the truth about the odds, and the second is to be clear about why the odds are what they are. Neither court denies review because it thinks the decision below was right. Both deny because review is a scarce institutional resource allocated on criteria that are mostly not about this case at all.

A lawyer who understands that writes a different document from one who does not.

What certiorari is for

28 U.S.C. § 1254 gives the Supreme Court discretionary jurisdiction over cases in the courts of appeals; § 1257 covers final judgments of state courts of last resort where a federal question is presented. The time limits sit in § 2101 and in the Court's own rules — ninety days from entry of judgment or denial of rehearing for a civil case, extendable by a Justice for good cause, but the ninety days is the operative deadline and the extension is not routine.

The Court's own Rule 10 tells petitioners what it is looking for, and it is worth reading as the instruction it is rather than as boilerplate. It states that review is "not a matter of right, but of judicial discretion," and that a petition "will be granted only for compelling reasons." The listed considerations are:

A conflict among the courts of appeals on an important federal question.
A conflict between a court of appeals and a state court of last resort.
A court of appeals decision on an important question that should be settled by the Supreme Court, or that conflicts with the Court's own decisions.
A state court of last resort deciding an important federal question in conflict with another state court of last resort or a court of appeals.
A departure from the accepted and usual course of judicial proceedings so far as to call for supervisory power.

And Rule 10 says the thing petitioners least want to hear: "A petition for a writ of certiorari is rarely granted when the asserted error consists of erroneous factual findings or the misapplication of a properly stated rule of law."

The conclusion that follows is that a petition arguing that the court below got it wrong will be denied, however right the argument is. The petition must argue that the question needs answering, by this Court, now, and that this case is the right vehicle for answering it.

The conflict, and what makes it real

A genuine conflict is the most reliable basis for a grant, and most asserted conflicts are not genuine.

What a real conflict looks like:

Squarely on the same legal question, not on a related question or a different application of the same standard.
Outcome-determinative — the case would have come out differently in the other circuit. If the results would be the same, there is no conflict worth resolving.
Acknowledged, ideally, by one of the courts. A circuit that expressly declines to follow another is the best evidence available.
Entrenched — the losing circuits have had the opportunity to correct themselves and have not. A conflict that arose last month may be left to percolate.
Recurring — the question arises often enough to matter, and the conflict produces different results for similarly situated parties.
Clean — the decisions can be described without heavy qualification.

What is not a real conflict, though petitions routinely claim otherwise:

Different outcomes on different facts under the same legal standard.
Tension in dicta.
A district court split, which is not a circuit conflict at all.
A conflict with an old decision the circuit has since limited.
A conflict that exists only if one reads a decision in a way that court has not.
A conflict the petitioner constructed by characterizing the decisions at a level of generality neither court used.

The respondent's brief in opposition will attack the conflict on each of these grounds, and the Court's law clerks are practised at spotting a manufactured one. A petition that overclaims a conflict loses credibility on everything else it says.

Vehicle problems

More petitions with genuine questions are denied for vehicle problems than for any other reason. A vehicle problem is anything that would prevent the Court from reaching and resolving the question cleanly.

The recurring ones:

The question was not raised or preserved below. Fatal, almost always.
The court below did not decide it, or decided it in the alternative, so that the judgment rests on an independent ground.
An independent and adequate state ground supports the judgment.
Standing, mootness, or another jurisdictional defect.
Factbound application. The case turns on the record rather than on the legal rule.
Interlocutory posture. The Court prefers final judgments, and an interlocutory case may return with a different record or become moot.
The record is unclear on a fact the resolution would depend on.
The petitioner would lose anyway on an alternative ground the court below identified.
The question is presented alongside several others, so that the case cannot be resolved on the clean question alone.

The practical consequence for a petitioner is that case selection matters more than brief writing. Counsel with a portfolio of cases raising the same question should petition in the one with the cleanest vehicle, not the one with the most sympathetic client or the largest judgment.

For a respondent, the brief in opposition should lead with vehicle problems, because they are the easiest ground for a denial and require no view about whether the question is important.

Anatomy of a petition

The question presented is the most important page. It appears first, it is what the Justices and clerks read first, and it frames everything.

Rules for writing it:

One question, or at most two. A petition presenting four questions signals that none of them is the point.
Short. Under fifty words is a good target; some of the best are under twenty.
Neutral in form, persuasive in substance. A question loaded with argument reads as advocacy and invites the respondent to say the question is not the one the case presents. But a genuinely neutral question can still be framed so that the answer seems obvious.
It should contain the conflict. "Whether, as the Second and Ninth Circuits have held but the Fifth Circuit has rejected, ..." tells the reader the whole story in one line.
It must match what the court below actually decided. A question the court did not answer is a vehicle problem stated on page one.

The other required components: parties to the proceeding and corporate disclosure; table of contents and authorities; opinions below; jurisdiction; constitutional and statutory provisions involved; statement of the case; reasons for granting the petition; conclusion; and an appendix containing the opinions below, the order denying rehearing, and the relevant provisions.

The statement of the case should be short, scrupulously accurate, and organized around the legal question rather than the equities. A statement that reads as a grievance is a statement that will be checked against the record.

The reasons for granting should be ordered: (1) the conflict, described court by court with the holdings quoted; (2) the importance and recurrence of the question; (3) why this case is a good vehicle — addressing the problems before the respondent raises them; and only then (4) why the decision below is wrong, briefly. That last section is the shortest and the least important, which is the opposite of how most petitions are written.

Length. The Court's rules cap a petition at 9,000 words. The best petitions are considerably shorter. A 5,000-word petition that says one thing clearly outperforms a 9,000-word petition that says six things.

The brief in opposition

A respondent has three choices, and the choice itself is strategic.

Waive. Permitted, and sensible where the petition is plainly meritless and the case is unremarkable. A waiver signals confidence. Its risk is that if the Court is interested, it will call for a response, and a brief filed after a call for response is filed under a spotlight — the Court has already noticed the case.

File a brief in opposition. The default in any case where the petitioner is competent and the question is real.

File and confess. Rare, and appropriate where the government or another party agrees that a conflict exists and should be resolved.

How to write the opposition:

Lead with the vehicle problems. They are the cheapest ground for denial and require no institutional judgment about the question's importance.
Dismantle the conflict. Take each case the petitioner cites and show that it is distinguishable, superseded, or decided on other grounds. Where a conflict is shallow, say that the question is percolating and would benefit from further development.
Deny the importance. The question arises rarely; the statute has been amended; the industry has adapted; the rule below is the settled majority position.
Defend the merits last, and briefly. The merits matter least at this stage, and a long merits defence signals that the earlier arguments were weak.
Correct the statement of facts where it matters, precisely and without heat. An opposition that fights every characterization loses the reader.

Length and tone. The rules permit 9,000 words. Most good oppositions are half that. The register is calm and institutional: this case does not warrant the Court's attention, and here is why — not: the petitioner is wrong and behaving badly.

Reading the signals

Between distribution and disposition, a petition emits signals that experienced counsel watch.

A call for a response (CFR) where the respondent waived means at least one chambers is interested. It is not a grant, but it is a substantial upgrade in the odds.

A relist — the petition is distributed for a second or subsequent conference — means something is happening. It may mean a Justice is writing a dissent from denial; it may mean a summary reversal is being drafted; it may mean the Court is holding for a related case. A single relist substantially increases the likelihood of a grant relative to the base rate; multiple relists more often precede an opinion respecting the denial.

A CVSG — a call for the views of the Solicitor General — occurs in cases with a federal interest and adds many months. The government's recommendation is highly, though not perfectly, predictive.

A hold — the petition sits undistributed or is repeatedly rescheduled while a related case is pending — usually ends in a GVR: grant, vacate, and remand in light of the intervening decision. A GVR is not a merits ruling, but it can be worth a great deal, because it returns the case to the court below with an instruction to reconsider.

A summary reversal without argument is rare and reserved for clear errors, most often where a lower court has disregarded the Court's own precedent.

Rescheduling before the first conference is common and means little on its own.

Amicus support at the certiorari stage

Amicus briefs at the petition stage do something quite specific: they establish that the question matters to people other than the petitioner.

What makes an amicus brief useful:

The amici are real and varied. Five trade associations from different industries, or a bipartisan group of former officials, or a set of states, says more than a single interested party.
They supply information the parties cannot. How often the question arises, what it costs, how businesses have structured around the conflict, what the practical consequences of each rule are.
They do not repeat the petition. A brief that restates the petitioner's argument is worse than no brief, because it wastes the Court's attention.
They are filed on time. Amicus briefs supporting a petitioner are due within thirty days of the docketing of the petition or the call for a response, and late briefs are simply not considered.

What to avoid: briefs from entities that appear to have been organized by the petitioner and have no independent interest; briefs that argue the merits of the underlying dispute; and volume for its own sake. Fifteen thin briefs are less persuasive than three substantial ones.

Coordination. Where a coordinated campaign is appropriate, someone should own it: identifying candidate amici, allocating topics so that briefs do not duplicate, ensuring counsel consent is obtained, and tracking the filing deadline. This work takes weeks and cannot begin after the petition is filed.

Rehearing en banc

Federal Rule of Appellate Procedure 35 governs, and its opening sentence is the one to internalize: en banc hearing or rehearing "is not favored and ordinarily will not be ordered."

The rule identifies two grounds:

(1) Necessary to secure or maintain uniformity of the court's decisions — an intra-circuit conflict. The panel decision cannot be reconciled with an earlier published decision of the same circuit.

(2) The proceeding involves a question of exceptional importance — including, the rule says, a case involving an issue on which the panel decision conflicts with the authoritative decisions of other courts of appeals.

The intra-circuit conflict ground is the stronger one, and it is the ground most often available. Most circuits follow a rule that a panel cannot overrule a prior published panel decision — only the en banc court can. So a petition establishing that the panel did exactly that is not asking the court to reconsider a judgment call; it is identifying a structural problem only the full court can fix.

How to establish it: quote the holding of the earlier decision and the holding of the panel, side by side, and show that they cannot both be law. Do not paraphrase. Do not argue at a level of generality that obscures the conflict. If the two decisions can be reconciled by any reasonable reading, the ground fails.

The exceptional importance ground is harder and requires showing that the question transcends this case: it affects a large number of cases, it involves a constitutional question of general application, it creates a circuit conflict, it has substantial economic or governmental consequences, or it concerns the administration of justice in the circuit.

Rule 35's certificate. The petition must begin with a statement that, in counsel's judgment, one of the grounds exists, expressed in the specific terms the rule requires. It is a formality, and omitting it is an unnecessary error.

Length. 3,900 words. This is a genuine constraint and it is the right one — an en banc petition is not a place for a comprehensive merits argument.

Panel rehearing, and why it is usually a different document

FRAP 40 governs panel rehearing, and it asks for something quite different from Rule 35. A petition for panel rehearing must "state with particularity each point of law or fact that the petitioner believes the court has overlooked or misapprehended."

That is the whole ground. Not that the panel was wrong on a question it considered — that it overlooked or misapprehended something: an argument the briefs made that the opinion does not address, a controlling authority not cited, a material fact in the record stated incorrectly, a holding broader than the question presented, or an internal inconsistency in the opinion.

Good panel rehearing petitions are short and specific. Two pages identifying one thing the opinion got factually wrong, with a record citation, is a better petition than fifteen pages re-arguing the appeal.

The most useful function of a panel rehearing petition is often not reversal but narrowing. A panel that has written more broadly than necessary will sometimes revise the opinion to limit its reach, which can matter enormously for a party's other cases and for an industry. A petition that says "we do not ask the Court to change the result, but the opinion's statement at page 14 sweeps beyond the question presented and will govern cases not before the Court" is asking for something a panel can give without conceding error.

Combining the two. Most practitioners file a combined petition for panel rehearing and rehearing en banc, which is permitted and efficient. But the two sections should be genuinely different: the panel section is particular and narrow; the en banc section is institutional and general. A combined petition that makes the same argument twice serves neither purpose.

Timing. Fourteen days after entry of judgment in a civil case without the United States as a party; forty-five days where the United States is a party. The deadline is jurisdictional in effect and extensions are granted sparingly and must be sought before it expires.

The effect on certiorari timing. A timely petition for rehearing suspends the finality of the judgment, and the ninety-day certiorari clock runs from the denial of rehearing rather than from the original judgment. This is the principal tactical reason to file a rehearing petition even where the prospects are slim — though a petition filed solely for delay is transparent and does the client no credit.

How en banc actually happens

Understanding the mechanics explains why the petition is written the way it is.

In most circuits, a petition for rehearing en banc is circulated to all active judges. Any active judge may call for a vote. If no judge calls for a vote, the petition is denied without any judge formally considering the merits at length. If a vote is called, a majority of the active, non-recused judges must vote in favour.

The practical consequences:

The petition's audience is not "the court." It is the one judge most likely to care. A petition that identifies a problem a particular judge or wing of the court has previously flagged has a mechanism for succeeding that a merely well-argued petition does not.
Judges also call for votes sua sponte, sometimes in cases where no petition was filed. A dissent from the panel decision materially increases the chance of this.
A panel dissent is the single most valuable asset an en banc petitioner can have, and it should be used: quoted, relied on, and treated as the statement of the problem.
Denials are sometimes accompanied by opinions dissenting from the denial. These are consequential beyond the case: they frame the issue for a certiorari petition, and the Supreme Court reads them.
Recusals matter. In a closely divided court, the recusal of two judges can change the arithmetic.
Circuit size matters. The largest circuit uses a limited en banc court rather than the full bench, which changes the calculation considerably.

The relationship between the two doors

En banc and certiorari are sequential, and the choices made at the first affect the second.

Filing en banc first is usually right where an intra-circuit conflict exists or where the panel opinion contains a factual or analytical error a panel could fix. It costs a few weeks and it preserves the certiorari clock.

It can also help the certiorari petition. A denial of rehearing en banc over a written dissent by several judges is powerful evidence that the question is important and the circuit is divided. Some of the strongest certiorari petitions are built around exactly that.

But it can hurt. A unanimous denial of rehearing, with no dissent, tells the Supreme Court that no judge of the circuit thought the case warranted further review. And a petition that changes the question presented between the en banc stage and the certiorari stage invites the argument that the new question was not pressed or passed on below.

The practical rule: frame the question the same way at both stages, file the en banc petition unless there is a reason not to, and treat any dissent from denial as the centrepiece of the certiorari petition that follows.

Worked example one: the split petition

Marguerite Vasquez represents a defendant that lost in the Eleventh Circuit on the scope of a federal statutory safe harbour. Three circuits read the safe harbour broadly; the Eleventh Circuit and one other read it narrowly.

Case selection. Her firm has two cases raising the question. The other has a larger judgment and a more sympathetic client, but the court there resolved the case on an alternative ground as well. She petitions in this one, because the judgment rests solely on the safe harbour question.

The question presented, after eleven drafts, runs twenty-eight words and names the conflict.

The conflict section takes each of the five circuits in turn, quotes the operative holding, and shows what result each rule produces on the same facts. She includes a table. She does not overclaim: one of the five decisions is arguably distinguishable, and she says so, explaining why the conflict is real without it.

The vehicle section comes before the merits and addresses three problems the respondent will raise: preservation (with record citations), the interlocutory question (the judgment is final), and an argument that the case is factbound (it is not, because the panel decided the legal question as a matter of statutory construction).

The merits section is 900 words.

Amicus support. Four briefs: a trade association, two companies in different industries that structure transactions around the conflict, and a group of law professors on the statutory text. Coordinated so that each covers different ground, all filed within the thirty-day window.

The disposition. The Court calls for the views of the Solicitor General. Ten months later the government recommends a grant. The Court grants.

What did the work was not the merits argument. It was case selection, an honest conflict description, and vehicle problems addressed before the respondent could raise them.

Worked example two: the opposition

Aleksander Nowak represents the respondent in a petition asserting a three-way circuit conflict on a question of arbitration procedure.

The first decision is whether to waive. He does not, because the petitioner's counsel is experienced and the conflict claim is superficially plausible. A waiver here would likely produce a call for a response, filed under scrutiny.

The structure of his opposition:

Vehicle, first. The petitioner did not raise the question in the district court and raised it in the court of appeals only in a footnote of a reply brief. The panel's opinion does not address it. Aleksander documents this with record citations and quotes the panel's actual holding, which rests on a waiver ground the petition does not mention.

The conflict, second. Two of the three cited decisions predate a statutory amendment. The third involved a materially different clause. He shows, case by case, that the results would be identical in every circuit on these facts.

Importance, third. The question arises in a narrow category of agreements; the drafting practice that produced the dispute has changed; and no court has been troubled by the issue in six years.

Merits, last, in two paragraphs.

Length: 4,100 words against a 9,000-word limit.

Disposition: denied without a relist.

The lesson. The strongest opposition argument was that the panel did not decide the question. That is a vehicle argument, it required no view about the conflict, and it was available because Aleksander read the opinion against the petition rather than against his own memory of the case.

Worked example three: the en banc campaign

Desmond Achebe's client lost 2–1 in a panel decision holding that a state-law claim was preempted. The dissent is vigorous and reads, in places, like an invitation.

The intra-circuit conflict. Desmond identifies a published decision of the same circuit from six years earlier that reached the opposite conclusion on materially indistinguishable facts. The panel opinion does not cite it.

The petition opens with the Rule 35 certificate and then, on page two, sets the two holdings side by side in a table: the earlier decision's language, the panel's language, and a one-line statement that they cannot both be law. That table is the petition.

The remainder — under 2,500 words — establishes that the question recurs (thirty-one district court decisions in the circuit in five years, cited), that the panel decision creates a circuit conflict as well, and that the panel's failure to address the earlier decision means the conflict was never considered.

The dissent is quoted three times and is the spine of the argument.

Amicus support at the en banc stage — permitted, and used here: two briefs, one from a state attorney general and one from an industry group, both addressing the practical consequences within the circuit rather than the merits.

The disposition. A judge calls for a vote. The court grants rehearing en banc, vacates the panel opinion, and, eight months later, reverses.

What made it work was that the argument was structural rather than substantive. The court was not asked to decide that the panel was wrong; it was asked to resolve an inconsistency that only the full court could resolve. That is the argument Rule 35 was written for.

State court petitions and the independent state ground

Certiorari from a state court of last resort under 28 U.S.C. § 1257 carries an additional obstacle that catches practitioners who work mainly in the federal system.

The independent and adequate state ground doctrine. The Supreme Court will not review a state court judgment that rests on a state law ground independent of the federal question and adequate to support the judgment. If the state court decided the case on state constitutional grounds, or on a state procedural rule, the federal question is advisory and the Court lacks jurisdiction.

Ambiguity is resolved by a presumption. Where a state court decision fairly appears to rest primarily on federal law, or to be interwoven with federal law, and the adequacy and independence of any possible state ground is not clear from the face of the opinion, the Court will accept jurisdiction — unless the state court includes a plain statement that its decision rests on separate, adequate, and independent state grounds.

What this means in practice:

For a petitioner, the opinion below must be read closely for any state law ground, and the petition must address it directly. Where the state court relied on parallel state and federal provisions, the petition should show that the state court treated the federal question as controlling.
For a respondent, the plain-statement rule is a reason to ask the state court, on rehearing, to clarify that its decision rests on independent state grounds. That single sentence can end the case.
State procedural bars — failure to preserve, untimeliness — are adequate state grounds if regularly applied. A petitioner facing one must show that the rule is not firmly established or is applied inconsistently.

"Final judgment" under § 1257 has been construed with some flexibility, and there are recognized categories in which a decision is treated as final though further state proceedings remain. Where the posture is unusual, address finality in the petition rather than hoping it goes unnoticed.

The economics and the client conversation

Both doors are expensive relative to their success rate, and clients deserve a candid account before the work begins.

What a certiorari petition costs. A properly prepared petition — case selection, conflict research across every circuit and relevant state court, multiple drafts of the question presented, vehicle analysis, appendix preparation, and printing — is a substantial engagement. Amicus coordination adds to it. Specialist appellate counsel typically charge for the petition as a discrete matter.

What the odds are. Base rate in the low single digits for paid petitions. A genuine, acknowledged, entrenched conflict on a recurring question in a clean vehicle raises that materially — but the honest number is still well short of even.

What the value is, apart from a grant. A petition can preserve a settlement position, signal to an industry that the question is live, generate a dissent from denial that shapes later litigation, or set up a GVR if a related case is decided. Those outcomes have value, and they should be identified explicitly rather than offered as consolation afterward.

When to advise against. Where the question was not preserved. Where the judgment rests on an alternative ground. Where the asserted conflict is manufactured. Where the client's real objective is delay — the cost is disproportionate and the professional exposure is real.

The en banc calculation is different. The petition is short, the cost is a fraction of a certiorari petition, the odds are better where a genuine intra-circuit conflict exists, and it preserves the certiorari clock. For most losing appellants with a colourable Rule 35 ground, it is the right recommendation.

Bringing in appellate counsel

A brief and unglamorous point that improves outcomes more than most drafting advice.

Bring specialist appellate counsel in early — ideally before the appeal, certainly before the rehearing petition. Trial counsel who have lived with a case for four years know it better than anyone and are, for that reason, systematically poor at seeing which issue is the one an appellate court will care about. The instinct to preserve everything and argue everything is right at trial and fatal on appeal.

What fresh counsel adds: ruthless issue selection; a reading of the opinion below as a stranger would read it; knowledge of the conflict landscape and of what the Court has recently taken; and the ability to tell the client that the strongest argument at trial is not the argument that gets certiorari.

What to avoid: replacing trial counsel rather than adding to the team. The record knowledge is irreplaceable, and the best appellate work is done by a small team combining both.

On the question presented specifically: have someone who has not worked on the case read the question cold and say what they think the case is about. If their answer differs from the intended one, the question is not finished.

A short list of avoidable errors

Missing the deadline. Ninety days for certiorari, running from denial of rehearing if a timely rehearing petition was filed; fourteen days for panel rehearing and en banc in an ordinary civil case. Calendar both the day the opinion issues, and calendar the extension application deadline separately.

Presenting four questions. It signals that none of them is the point, and it creates a vehicle problem — the Court cannot resolve one question cleanly if the case comes wrapped in others.

A question the court below did not decide. Page one, and the case is over.

Overclaiming the conflict. The clerk who checks one of the cited cases and finds it distinguishable will discount everything else the petition says.

Leading with the merits. The most common structural error, and it tells the reader that counsel does not know what the document is for.

Ignoring the vehicle problems. They will be raised. Better to address them.

Re-arguing the appeal in a rehearing petition. Rule 40 asks what the panel overlooked or misapprehended. A brief that simply disagrees is denied without discussion.

Omitting the Rule 35 certificate. Free points, lost.

Failing to use a panel dissent. It is the single most valuable asset an en banc petitioner has.

Amicus briefs that repeat the petition. Worse than none, because they consume attention without adding information.

Late amicus briefs. Simply not considered.

Changing the question between stages. It invites the argument that the new question was not pressed or passed on below.

What the whole exercise is really about

It is worth stating plainly what a lawyer is doing when they file one of these documents.

They are not asking a court to correct an error. They are making a case that an institution with a finite capacity to decide should spend some of that capacity here rather than somewhere else. The Supreme Court will decide perhaps sixty cases this Term out of many thousands of federal appellate decisions. A court of appeals sitting en banc will use that machinery a handful of times a year. Every grant is a choice not to hear something else.

That framing changes what the document says. It stops being about the client's grievance and starts being about the state of the law: here is a question that arises constantly, on which the courts have divided, producing different outcomes for identically situated parties, which the lower courts cannot fix themselves, presented here in a case that raises it cleanly and nothing else.

Petitions written that way still usually fail, because the numbers are what they are. But they fail for the right reasons, and the ones that succeed are almost always the ones that understood the question they were actually answering.